[
https://issues.apache.org/jira/browse/SOLR-4733?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13673764#comment-13673764
]
Mark Miller commented on SOLR-4733:
-----------------------------------
bq. we can improve rollback support by dropping all the internal tlog state and
forcing open a new realtime reader
Okay - I can open a JIRA issue for that then. Because as it is, it seems we
have to say anything that uses the update log does not support rollback -
SolrCloud or not.
bq. Rollback is never going to be a "first class" feature... there's no
transaction isolation
I don't really follow that - commit is the same way, so it seems rollback is
just as first class in that regard - it's something you have to understand, but
rollback is still simply the same as commit to me - you have one global
transaction to work with.
> Rollback does not work correctly with tlog and optimistic concurrency updates
> -----------------------------------------------------------------------------
>
> Key: SOLR-4733
> URL: https://issues.apache.org/jira/browse/SOLR-4733
> Project: Solr
> Issue Type: Bug
> Affects Versions: 4.3, 4.2.1
> Environment: Ubuntu 12.04.2 LTS
> Reporter: Mark S
> Labels: solrj
> Attachments: SOLR-4733.patch
>
>
> When using the updateLog, attempting to rollback atomic updates still causes
> post-rollback atomic updates to still report a conflict with the version
> assigned to the update posted prior to the rollback
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
